Updates to run "base" s2s-prototype4 #80
Conversation
making readme consistant with user.yaml.default
…i cplwav model_config block. Reinstating config.wave block in JGLOBAL_FORECAST. Pointing EXECwave to HOMEgfs/exec directory for WW3 util executables (changed link_fv3gfs.sh accordingly). Removing debug options from compile.sh line in build_fv3.sh.
…coupled-crow-addWW3jobs
…coupled-crow-addWW3jobs
…coupled-crow-addWW3jobs
…coupled-crow-addWW3jobs
IC for s2s model and made it platform dependent. (Orion location is temporary and needs to be replaced, can also be changed in cases/coupled_free_forecast.yaml to users prefered location)
copying suite xml file to rundirectory
post jobs for atmosphere are no longer triggered on cold starts
Not sure if other jobs will now fail though or if what is in your .bashrc file will effect the success/failure of the run.
|
@JianKuang-NOAA I wanted to add Lydia @LydiaStefanova-NOAA and Jiande @jiandewang as reviewers but could not. |
JessicaMeixner-NOAA
changed the title
Neither of them are in the NOAA-EMC org member list...that may be why. I have sent them both invites to join the org. |
|
@JessicaMeixner-NOAA This PR contains J_* scripts and build/link updates, along with some overlaps with my work. I think this should be merged first before my updates, which may be more Orion-specific. |
|
@JianKuang-NOAA, in reference to the issue @LydiaStefanova-NOAA saw, is there a way to purge a users modules before the forecast job runs or do we need to add a note in the README that users might need to purge modules if conflicts arise and cause the forecast job to fail? |
|
@JessicaMeixner-NOAA COuld you please (re)send me an email about the issue that you are talking about? In general, user-loaded modules shouldn't affect the job because the modules are handled by ush/load_fv3gfs_modules.sh script. I need to get more in-detail description about the issue. |
|
Jian - here it is:
An issue I ran into so far with testing the new workflow on Orion:
For my first attempt, I started from a window which had some previously
loaded modules, and gfs.forecast.medcold failed. Among the error messages
was a complaint about "***HDF5 library version mismatched error***" - and
one of the modules I had previously loaded was indeed hdf5/1.10.5.
Trying the same case with a clean slate (no modules loaded except for
contrib, rocoto, and intelpython3) works just fine/
The corresponding logs are
…-- for the case with errors (run with previously loaded modules):
/work/noaa/marine/lstefano/scratch/Jun2Test/COMROOT/test_3d/logs/2013040100
-- for the case that is running ok so far (run with no previously loaded
modules):
/work/noaa/marine/lstefano/scratch/Jun2Test/COMROOT/test_3d_fresh/logs/2013040100
(The model itself is built in
/work/noaa/marine/lstefano/scratch/Jun2Test/global_workflow)
|
|
@JessicaMeixner-NOAA @LydiaStefanova-NOAA I found the issue: The following lines of script ush/load_fv3gfs_modules.sh do not work: Find module command and purge:source "$HOMEgfs/modulefiles/module-setup.sh.inc" The fix is easy. I will include it in my later PR. How's that sounds like? |
|
@JessicaMeixner-NOAA Could we merge it now or still wait for some testings? |
|
@JianKuang-NOAA there are still some issues with ocean-post, but we could probably merge it now and then fix those as a next step. the forecast and atm post should work without issues on both hera and orion. |
|
Should we change OCN_FHRLST: !calc "tools.seq(dimval.fhr, dimval.fhr+doc.output_settings.OCN_INTERVAL, 6)" to OCN_FHRLST: !calc "tools.seq(dimval.fhr, dimval.fhr+doc.output_settings.OCN_INTERVAL-6, 6)"? because there is one cycle was double processed (see below) when we do ocean post. |
* Update FV3 submodule * Regression test log files on Hera * Regression test log files on WCOSS Cray * Regression test log files on WCOSS Dell * Update FV3 submodule
Update hello.yaml
This PR is for running the "base" case of prototype 4. This means checking out and building the updated tag of ufs-s2s-model s2s_prototype4.0, running with CCPP (suite file FV3_GFS_v15p2_coupled), updated oro data with lakes (using new fix file directories), but not running with wave coupling. Other updates include:
This does not include:
Before this PR is accepted, someone else should test that they can replicate on hera that everything works including the updated tar file archiving structure and on orion that they can run the IC, forecast job and atm posts. I'm slightly worried that what someone has in their .bashrc on orion will interfere with successfully running forecast/atm post on orion.